We develop an instrument-free method to correct for endogeneity when estimating price responses and report an analytics tool we developed for a hotel chain to estimate the price response functions in aggregate level. Our approach overcomes the issues of price endogeneity, censored demand, and unobservable demand shocks in the real data using a structural eqaution modeling. To handle these issues, we model a hotel’s revenue management problem as a single-period revenue maximization problem with a stochastic and price-dependent demand. With a piece-wise linear approximation to the loss function due to capacity limit, the revenue function becomes a piece-wise quadratic function of the price. Under a rather general condition, we derive a closed-form formula for the best price to this revenue function, which leads to a closed-form log-likelihood function for the observed sales and price. With a global optimization routine we find the optimal parameter values for our data. The simulation study shows that the more segments in the piece-wise linear approximation, the more accurate the approach is. Especially, the approach with odd segments respectively performs better than that with even segments.